AuditSpine
Pipeline Proof Status
Last updated: 2026-04-05
Proof pipeline registry
Pipeline Customer Dataset Tier 1 Tier 2 Tier 3
NYC Taxi 2024 VIRIDE 7,667,792 rows
PASS
PASS
PASS
NYC Taxi Longitudinal VIRIDE 2,964,624 rows · day-by-day temporal simulation
PASS(300K)
PASS(2.87M)
NOT-PLANNED
IEEE-CIS Fraud VIRIDE 50,000 rows (proof subset)
PASS
PASS(50k)
PENDING
Supermarket Sales 66DEGREES 1,000 rows baseline · 300,000 rows scale
PASS
PASS(300K)
PASS
NYC Taxi Yellow 2024-01
VIRIDE — 7,667,792 rows — all tiers complete
Run metadata — NYC Taxi
Run IDc389e830-48aa-45b0-a0e9-5eabbf8d4c3b
CustomerVIRIDE
DatasetNYC Taxi Yellow 2024-01 — 7,667,792 rows
Pipeline versionv2.1.5
RMSE gate3.9177 / gate 4.0 — PASS
Dataset SHA-256abee0ee30bba9aa405ec2633bce980900549ac7b42ea53aa6af6ff14e50d56a6
↑ back to top
NYC Taxi Longitudinal 2024-01
VIRIDE — 2,964,624 raw rows — day-by-day temporal simulation — >1M silver proven — Dataflow not planned
Run metadata — NYC Taxi Longitudinal (nyc_scale_20260413T222003Z)
Run IDnyc_scale_20260413T222003Z
Bronze SHA-256C4D59DA7BBC8ABAEEEB1727947EE93D9891A71ACB42854BD80DB1571B2030510
DatasetNYC TLC Yellow Taxi 2024-01 — 2,964,624 raw rows — 35 day slices
Cap 10K/day300,348 silver rows — 21.9s
Cap 50K/day>1M threshold MET — 1,510,429 silver rows — 31.9s
Uncapped2,869,714 silver rows — 44.0s
Cross-grain audit (uncapped)daily silver sum == monthly silver ref — delta $0.00 — PASS
Quality exclusion (raw → silver)$53,882,224.76 raw → $53,087,675.11 silver — delta $794,549.65 — intentional, closed — fare ≤ 0 or distance ≤ 0
Day sealsSHA-256 per day slice (35 sealed day records)
Scale pathGCP Dataflow at >10M rows — same Bronze→Silver→Gold logic — no rewrite
↑ back to top
IEEE-CIS Fraud Detection
VIRIDE — 590,540 rows (full dataset) — AUC 0.9006 — T3 pending
Run metadata — IEEE-CIS Fraud
Run IDieee-cis-t031-v2-20260406
CustomerVIRIDE
DatasetIEEE-CIS Fraud Detection — 590,540 rows (full Kaggle dataset)
Pipeline versionv2.1.5
ML AUC-ROC0.900552 — gate 0.897 — PASS
Features76 (36 baseline + 40 V-columns)
Dataset SHA-2563a5c83ab6b3cc13dcabe5ffa9f522307fd5f7f7b6e6f6a60c32284ca6283d642
↑ back to top
Supermarket Sales Pipeline PoC
66DEGREES — 1,000 rows baseline / 300,000 rows scale — all tiers complete
Run metadata — Supermarket Sales (Pass 1 / baseline)
Bronze SHA-256 (cross-tier)901AA9D1999DF4C620B547A34B250912610A04342686C62CC2DFB82B163812A8
Cross-tier parityT1 / T2 / T3 Bronze SHA identical — PASS
DatasetKaggle Supermarket Sales — Jan–Mar 2019 — 1,000 rows — 3 branches
SchemaStar schema: dim_branch (3) · dim_product (6) · fact_sales (1,000)
Pass 2 cross-grain auditJan + Feb + Mar = $322,966.75 — delta $0.00 — PASS
Scale run metadata — Tier 2 · 300× (scale_20260413T213226Z)
Run IDscale_20260413T213226Z
Bronze SHA-2564DEF6696BD4FB7DEEB42AA6CD0553E4D8131B52703ADBBC57B1BE461DB017405
Config sealCD530B89B04C8559...
Rows processed300,000 (300× baseline) — Apache Spark 4.1.1 / Docker
Cross-grain audit (scale)$92,794,944.40 — delta $0.00 — PASS
Wall time131.9s (~2m 12s) — single-node silverFoxDev
Gold report54 rows (3 cities × 6 products × 3 price tiers) — all 4 window functions
Scale pathDataflow at >1M rows — same Bronze→Silver→Gold logic — no rewrite
↑ back to top
Pass
Running
Queued / Pending
Warning
Failed